13. AMINO-ACIDS AND PEPTIDES By H. D. Law (Liverpool Regional College of Technology Byrom Street Liverpool 3) SUPERFICIALLY, polypeptides resemble polymers to such an extent that many chemists working in other areas must have wondered at the seeming com- plexity of peptide chemistry. Even when the twenty-fold variation in the monomer unit is taken into account the problem of peptide structure or synthesis seems to be an aggregate problem requiring the repetitious application of one process for its solution. To peptide chemists the similarity to polymers has often seemed remote. Each amino-acid has its own attendant difficulties ; no two peptide bonds are exactly alike. It is not difficult to discern the general strategy of structural elucidation or synthesis but its execution has usually depended on the exploitation of a battery of techniques each possessing its own general advantages and particular shortcomings.Progress in peptide chemistry has therefore been the sum of many diverse advances. In conse- quence it is inevitable that a review of this type will be somewhat fragmentary. However sufficient is now known about the idiosyncrasies of individual amino-acids and peptide bonds for the peptide chemist to begin to reconsider polypeptides as polymers. A recurrent theme in this review is the development of techniques which can be automated. Such techniques have now been described not just for amino-acid analysis but for residue sequence determina- tion and for polypeptide synthesis as well. These techniques will certainly be improved and may even be displaced by others yet to be devised but it is no exaggeration to say that their advent marks the beginning of a new phase in polypeptide chemistry.An end to the technical difficulties of studying the covalent structures of polypeptides and proteins may be in sight and with effective automated methods a statistically useful approach to the chemistry and biology of these molecules would become possible. Amino Acids.-The diversity of form of natural r-amino-acids has long ceased to cause surprise'. Most unusual amino-acids are non-protein in origin; plants and particularly fungi are rich sources of them. and each year gives rise to a new crop of compounds of this type. New amino-acids of fungal origin include trans-l-methyl-4-n-propyl-~-proline,~ ~-2-amino-3,3-dimethyl-aminopropionic acid,3 ~-2-amino-4,4-dichlorobutyric acid,3 2,3-dihydro-2- L.Fowden Ann. Reo. Biochem. 1964 33 173; L. Fowden D. Lewis and H. Tristram. iM.. 1967; 29 89. -B. J. Magerlein R. D. Birkenmeyer R. R. Herr and F. Kagan J. Amer. Chem. SOC.. 1967. 89. 2459. ' R. R. Herr D. J. Mason T. R. Pyke and J. F. Zieserl Biochemistry 1967,6 165. P* 452 H. D.Law 0x0-oxazol-5-yl glycine4 (l) y6-dihydro~yleucine,~and dehydroalanine.6 Linatine a vitamin-B antagonist in flax seed has been identified as the N'-y-L-glutamyl derivative of 1-amino-D-proline. Seeds of Cycas circinalis have yielded 1-amino-2-methylaminopropionic acid.8 OH OH IC.-CH,*R'/'R'-CH H I C -CH2 NH,R2 .-* /\R'-CH H R' I I CH-YH,I COT (1) iI R' ,K t CH,R'I ,CH ,G=CH -R' R'- C' ICH I CH R'- I CH *N 1 R2 *N I R2 ii i CH R' *OR' R2 R2 (2 ) (3) One protein elastin is proving to be a source of unusual amino-acids and it is clear that peptide chain cross-linking of a type not yet encountered elsewhere occurs in this protein.Lysinonorleucine [NE-(5-amino-5-carboxypentyl)-ly~ine],~ desmosine [2; R' = [CH,],-CH(NH,)*CO,H R2 = [CH2I4* CH(NH,).CO,H] isodesmosine [3; R' = [CH,] -CH(NH,).CO,H R2 = [CH,] -CH(NH,) CO,H R3= [CH,] -CH(NH,) CO,H] lo and now R. Reiner and C. H. Eugster Helv. Chim. Acta 1967 50 128; see also C. H. Eugster and T. Takemoto ibid. p. 126; H. Goth A. R. Gagneux C. H. Eugster and H. Schmidt ibid. p.137. ' T. Wieland and H. Wehrt Annalen 1966 700 120; P. Pfaender and T. Wieland ibid. p. 126; T. Wieland and V. Georgi ibid. p. 133; V. Georgi and T. Wieland ibid. p. 149; T. Wieland and U. Gebert ibid. p. 157; T. Wieland and J. X. de Vries ibid. p. 174. E. Gross and J. L. Morel] J. Amer. Chem. SOC. 1967 89 2791. H. J. Klosterman G. L. Lamoureux and J. L. Parsons Biochemistry 1967,6 170. * A. Vega and E. A. Bell Phytochernistry 1967,6 759. C. Franzblau M. Sinex B. Faris and R. Lampidis Biochem Biophys. Res. Comm. 1965 21 575. lo J. Thomas D. F. Elsden and S. M. Partridge Nature 1963 200 651; S. M. Partridge D. F. Elsden J. Thomas A. Dorfman A. Telser and H. Pei-Lee Nature 1966,209 399. Amino-acids and Peptides merodesmosine which is probably one of the isomeric structures [4;R' = [CH2]2*CH(NH2)*C02H have all been R2= [CH2],*CH(NH2)*C02H]," isolated from elastin.Tracer studies indicate that these amino-acids are derived from lysine residues which are involved in an oxidative cross-linking process when elastin fibres are formed from the soluble precursor protein. The process is postulated to involve aldehyde formation by oxidative de- amination of the E-amino-groups of lysine residues followed by stepwise condensation via aldol and Schiff base intermediates. In support of this hypo- thesis merodesmosin was obtained hydrolytically from young elastin which had been treated with alkali and then reduced with borohydride. Another reported new amino acid N*-(2-amino-2-carboxyethyl)ornithine, obtained from alkali-treated proteins,I2 could have been formed presumably in a similar way.An unusual amino-acid isolated from butterflies and previously thought to be the tetrahydro-4-oxoquinolinecarboxylic acid derivative (5),' is really 3-hydroxy-~-kynurenine (6).14 The error arose because of the ready cyclization of the monocyclic compound in the mass spectrometer. Several amino-acids including ~-cystathionine," ~-allocystathionine,~~ P-methyllanthionine,' ( +)-and ( -)-cis-S-(P-styry1)-L-cysteine S-oxides,' 2,3-dihydro-~-tryptophan,'~2,3-dihydro-5-hydroxy-~~-tryptophan,'~ various proline derivatives," and 3,3,3-trifl~oroalanine,~~ have been the subjects of recent synthetic studies. A general method for the preparation of P-perfluoro- alkylalanines2' (7; R = perfluoroalkyl) utilises the reaction between the 0 OH (5) B.C. Starcher S. M. Partridge and D. F. Elsden Biochemistry 1967,6 2425. K. Ziegler I. Melchert and C. Lurken Nature 1967,214,404. l3 K. S. Brown jun. J. Amer. Chem. SOC. 1965,87,4202. l4 T. Tokuyama S. Senoh T. Sakan K. S. Brown jun. and B. Witkop J. Amer. Chem. SOC. 1967,89 1017. M. L. Snow R. S. Dombro and C. Ressler J. Org. Chem. 1967,32 246. l6 H.-D. Belitz Tetrahedron Letters 1967 749. J. F. Carson and L. E. Boggs J. Org. Chem. 1967,32 673. 18 J. W. Daly A. B. Mauger 0.Yonemitsu V. K. Antonov K. Takase and B. Witkop Bio-chemistry 1967,6 648. l9 R. H. Andreatta V. Nair A. V. Robertson and W. R. J. Simpson Austral. J. Chem. 1967 20,1493; C.B. Hudson and A. V. Robertson ibid. 1967,20,1511,1521;M. Viscontini and H. Biihler Helu. Chim. Acta 1966,49 2524. 'O F. Weygand W. Steglich and F. Fraunberger Angew. Chem. Internat. Edn. 1967,6,808. W. Steglich H.-U. Heininger H. Dworschak and F. Weygand Angew. Chem. Internat. Edn. 1967,6 807. 454 H.D. Law NH NH Ac I (7) ii R&O * C(N2)*C0,Et R FC=C -CO Et I\ 0lcP (8) 1 Me (9) Reagents i CH,CN solution light ;ii H, AcOH PtO ; iii conc. hydrochloric acid 90" 18 appropriate perfluorocarboxylic anhydride and ethyldiazoacetate to form the perfluoroacyldiazoacetic ester (8). Photoaddition of this compound to acrylo- nitrile with displacement of nitrogen gives the 2-methyl-5-perfluoroalkyl-4-oxazole carboxylic ester (9),from which the amino-acid is obtained by catalytic hydrogenolysis followed by hydrolysis of the resulting N-acetyl P-perfluoro- alkyl alanine ester.The amino-acid can be resolved by way of the corres- ponding trifluoromethyl perfluoroalkyl oxazolone (10). Trifluoromethyl-2H-(1 0) oxazoiones react with amino-esters to give predominantly the diastereoisomers with the same absolute configuration at both asymmetric centres.22a. 23 Direct asymmetric synthesis of a-amino-acids has received considerable attention,24.25 although from a preparative standpoint methods dependent >7 -Proceedings of the Eighth European Peptide Symposium Noordwijk The Netherlands September 1966 ed. H. C. Beyerman A. Van De Linde and W. Maassen Van Den Brink North- Holland Publishing Co.Amsterdam 1967; (a)W. Steglich D. Mayer X. Barocio De La Lama H. Tanner and F. Weygand p. 67; (b) E. Scoffone A. Fontana F. Marchiori and C. Benassi p. 189; (c) Review E. Lederer and B. C. Das p. 131; (6)A. Prox and F. Weygand p. 158; (e) K. PoduSka H. Maassen Van Den Brink Zimmermannovii J. Rudinger and F. S6rm p. 38; (f) Brandenburg quoted by H. Zahn p. 43 ;(9)Th. Wieland and Chr. Birr p. 103 ;(h)L. Zervas I. Photaki. C. Yovanidis J. Taylor I Phocas and V. Bardakos p. 28; (i)M. Brenner p. 1 ;0')G. T. Young p. 55; (k) J. H. Jones B. Liberek and G. T. Young p. 15; (I) V. Gut reported by J. Rudinger p. 89; (m) H. C. Beyerman C. A. M. Boers-Boonekamp W. J. Van Joest and D. Van Den Berg p. 117;(n)H. Zahn T. Okuda and Y. Shimonishi p.108; (0)A. Patchornik M. Fridkin and E. Katchalski p. 91 ;(p) I. Z. Siemion and D. Konopinska p. 79; (4)R. B. Merrifield and A. Marglin p. 85; (r) H. Klostermeyer J. Halstrm P. Kusch J. Fohles and W. Lunkenheimer p. 113;(s)J. Beacham P. H. Bentley G. W. Kenner J. J. Mendive and R. C. Sheppard p. 235; (t) J. S. Morley p. 226; (u) M. Rothe I. Rothe T. Toth and K.-D. Steffen p. 8; (0)J.-P. Carrion B. Donzel D. Deranleay K. Esko P. Moser and R. Schwyzer p. 177. 23 F. Weygand W. Steglich and X. Barocio De La Lama Tetrahedron 1966 Suppl. 8 Part 1 p. 9. 24 K. Harada and K. Matsumoto J.Org. Chem. 1967,32 1794. 25 K. Harada Nature 1966,212 1571; J. Org. Chem. 1967,32 1790. Amino-acids and Peptides 455 on resolution are still to be preferred. The resolution of t-butyl N-trifluoro- acetyl-(& )-alaninate by g.1.c.over Chromosorb W coated with cyclohexyl N-trifluoroacetyl-L-valyl-L-valinateis a noteworthy development.26 Asym- metric transformations have also been used to determine the absolute con- figuration of amine~.~'.~~ Peptides Structural Elucidation.-Now that automated ion-exchange pro- cedures for amino-acid analysis are available this facet of structural elucidation presents no general difficulties but tryptophan-containing peptides owing to the instability of the indole moiety under acidic conditions do require special treatment. In a recently developed technique reaction of the peptide with o-nitrobenzenesulphenyl chloride in acetic acid solution quantitatively con- verts tryptophan residues into the 2-o-nitrophenylsulphenyl derivatives which are estimated spectroscopically [h,,,.280 and 365 mp E 16,700and E 4400].22b Difficulties arising from the presence of tryptophan in a peptide are often more than offset by the possibility of achieving selective peptide bond cleavage at this residue notably by treating the peptide with N-bromosuccinimide.29 Ozonolysis of tryptophan-containing peptides results in the conversion of the tryptophan residues into kynurenine residues3' Cleavage at these residues Reagents i electrolytic reduction ; ii H+/H20. (11) may be brought about by electrolysis followed by mild acid hydrolysis [(11)- (12)] although some cleavage at the kynurenine amino-group occurs.22b N-Bromosuccinimide may be used for the cleavage of tyrosyl peptide bonds but when tyrosine is N-terminal a 6-hydroxyindole derivative is obtained instead of the expected ~pirolactone.~ For the determination of N-terminal residue sequences the Edman method remains supreme.32 Phenyl isothiocyanate is condensed with the peptide and the N-terminal residue is cleaved from the resulting phenylthiocarbamoyl 26 E. Gil-Av and B. Feibush Tetrahedron Letters 1967 3345. 27 0.Cervinka Coll. Czech. Chem. Comm. 1965,30 1738. 0.Cervinka Coll. Czech. Chem. Comm. 1966,31. 1371. 29 B. Witkop Adu. Protein Chem. 1961 16,221. 30 F. M. Veronesse A. Fontana E. Boccu and C. A. Benassi Gazzetta 1967,97 321. 31 M. Wilchek T. F. Spande B. Witkop and G. W. A. Milne J. Amer. Chem. SOC. 1967 89 3349. 32 R.F. Doolittle Biochem. J. 1965,94 742; D. G. Smyth and D. F. Elliott Analyst 1964,89,81. 456 H.D. Law peptide as the 2-anilinothiazolin-5-one which rearranges to the 3-phenyl-2- thi~hydantoin.~~ An apparatus the ‘protein sequenator’ has been described in which the condensation and the cleavage of the thiazolinone are performed a~tomatically.~~~ 35 The various stages of the degradation are carried out in a spinning cup so that the solution is spread as a thin film on the surface of the cup ;air is excluded to reduce oxidative desulphuration of the thiocarbamoyl derivative. All solvents and reagents are carefully purified particularly to render them aldehyde-free so that no terminal amino-groups are blocked by side reactions. The formation of the thiazolinones is carried out under anhydrous conditions and the extracted thiazolinones are subsequently con- verted into the hydantoins in a separate step.In this way exposure of the peptide to hydrolysing conditions is avoided. More than fifteen residues can be removed by this technique every 24 hr.; the yield per cycle exceeds 98%. Approximately 0-25pmoles of protein are required. The sequence of the first sixty residues in apomyoglobin (from humpback whale) was determined successfully by this method which approximates to the theoretical limit if 2% loss is made at each step. Of course a small improvement in the efficiency at each step would extend quite considerably the length of sequence which could be determined. This possibility and the difficulty created by the detection on chromatograms of increasing amounts of phenylthiohydantoins from non-terminal amino-acids as the degradation progresses are being investigated and it may be expected that the applicability of the technique will be extended.Because of solubility difficulties this automated method is not likely to be readily applied to studies of small peptides and in preliminary investigations which might lead to an automated technique the pe~tide,~~? 37 or the cyclised residue,38 has been attached to a polymeric resin. A degradation similar to the Edman in which N-thiobenzoyl peptides are cleaved by treatment with tri- fluoroacetic acid has been de~cribed.~’This method seems particularly promising in view of the water-solubility of the reagent involved sodium thiobenzoylthioglycollate and of the detailed information already available on the c.d.properties of thiobenzoyl amino-acid~.~’ Dansylation4’ [reaction with 5-dimethylaminonaphthalene-1-sulphonyl chloride] and hydrolysis of the corresponding dansyl peptide to give the fluorescent dansyl derivative of the N-terminal amino-acid is a valuable alternative to the 2,4-dinitrofluorobenzene approach for the determination of N-terminal residues. The elucidation of the amino-acid residue sequence of 33 P. Edman Acta Chem. Scand. 1956 10 761; D. Bethell G. E. Metcalfe and R. C. Sheppard Chem. Comm. 1965,10 189. 34 P. Edman Thromb. Diath. Haernorrhug. 1963 Suppl. 13. 35 P. Edman and G. Begg European J. Biochem. 1967,1 80. 36 R.A. Laursen J. Amer. Chem. SOC. 1966,88,5344. 37 H. Jeschkeit R. Henkel and H. Lehmann 2.Chem. 1967,7 191. 38 G. R. Stark Fed. Proc. 1965,24 225. 39 G. C. Barrett Chem. Comm. 1967,487. 40 G. C. Barrett J. Chem. SOC.(C),1967 1. 41 W. R. Gray and B. S. Hartley Biochem. J. 1963,89 379; L. B. Smillie and B. S. Hartley ibid. 1966 101,232; W. R. Gray “Methods in Enzymology” Vol. 11 Ed. C. W. Hirs Amino-acids and Peptides 457 apamin (13) a C.N.S. stimulator for bee venom provides a recent example of the application of this technique.42 Dansylation has been used in conjunction with hydrazinolysis for C-terminal residue identifi~ation.~~ Peptides can be recovered from dansyl derivatives by reduction with sodium-liquid ammonia at -70°.44 T>e photolytic decomposition of 2,4dinitrophenyl amino-acids proceeds with the formation of 2-substituted 6-nitrobenzimidazole 1-oxides and 4-nitro-2-nitr xoaniline; the relative proportions of each are pH de- pendent4 Peptides in dilute aqueous solution at pH 7 react stoicheiometrically with the hydroxoaquotriethylenetetraminecobalt(m) ion to form the chelate [Co trien NH2CHR*COz]+ + (14) of th'e N-terminal residue.46 Reaction with the hydroxoaquobisethylenediaminecobalt(II1)ion to form the I I I I Cys .Asn .Cys .Lys .Ala .Pro.Glu Thr. Ala. Leu. Cys. Ala. Arg. Arg. Cys. Gln .Gln .His. NH HNCH-CO. NH .CH-CO. R* RZ I R' \ r/-N 1++ (1 4) Reagents [Co trien (H,O) (HO)] ++ rate-determining step 42 R. Shipolini A. F. Bradbury G.L. Callewaert and C. A. Vernon Chem. Comm. 19i7 679. 43 B. Mesrob and V. HoleySovskjr Coll. Czech. Chem. Comm. 1967,32 1976. 44 Z. Tamura T. Tanimura and H. Yoshida Chem. and Pharm. Bull. (Japan),1967,15,252. 45 D. J. Neadle and R. J. Pollitt J. Chem. SOC. (C) 1967 1764. 46 D. A. Buckingham J. P. Collman D. A. R. Happer and L. G. Marzilli J. Amer. Chem. SOC. 1967,89 1082. 458 H.D. Law [Co en 2NH2-CHR.C02]+ complex is apparently ~imilar.~' + Studies with dipeptides indicate that the complex formed most rapidly has the L-configura- tion at the cobalt atom48 when the C-terminal residue of the dipeptide is L and D-asymmetry at the cobalt atom when the C-terminal residue is D. In contrast to these results the alkaline hydrolysis of simple peptides in the presence of copper@) or nickel(I1) is reported to give selective release of the C-terminal residue; chelation protects four or five peptide bonds from hydr~lysis.~' Mass spectrometry has proved useful for the determination of the structures of a number of small peptides and depsipeptides but a limiting factor in this approach is the volatility of the sample.22" In principle however the method could be used to elucidate the structures of peptides obtained by partial hydrolysis from larger molecules and might therefore be developed to permit sequence studies of proteins.In a model study of this kind N-trifluoroacetyl peptide methyl esters obtained by methanolysis from a cyclic nonapeptide and purified by g.l.c. were identified by mass spectrometry; this permitted the sequence of residues in the nonapeptide to be deduced.22d R .co .NH -CHR~.co ...NH .c~~n-1.c& (1 6) Cleavage of the peptide bond [(l5) -+ (16)] probably proceeding in a step- wise manner from the C-terminal residue seems to be a fundamental mode of peptide degradation in the mass spe~trometer.~~-'~ Hence if the N-terminal residue is distinguishable e.0. because it possesses a characteristic acyl substituent the sequence may be deduced by discovering what mass additions to this unit would account for the spectrum observed. Other methods of fragmentation do occur in linear peptides [e.g. (15) -+ (17)]. In cyclodepsi- peptides the initial ring opening generally involves up to three basic routes [(18) -+ (19) (18) -+ (20) and (18) -+ (21)]; the significance of each depends 47 D.E. Allen and R. D. Gillard Chem. Comm. 1967 1091. 48 R. D.Gillard Chem in Britain 1967,3 205. 49 R. H. Andreatta H. C. Freeman A. V. Robertson and R. L. Sinclair Chem. Comm. 1967,203. 50 M.Barber P. JollCs E. Vilkas and E. Lederer Biochem Biophys. Res. Cornm. 1965 18,469. 51 N.S. Wulfson V. A. Puchkov B. V. Rozinov Yu. V. Denisov V. N. Bochkarev M. M. Shemyakin Yu. A. Ovchinnikov A. A. Kiryushkin E. I. Vinogradova and M. Yu. Feigina Tetra-hedron Letters 1965,2805. 52 F. Weygand A. Prox H. Fessel and K. Kun Sun 2.Naturforsch. 1965,206 1169. 53 N. S. Wulfson V. A. Puchkov V. N. Bochkarev B. V. Rozinov A. M. Zyakoon M. M. Shemyakin Yu. A. Ovchinnikov V. T. Ivanov A. A. Kiryushkin E. I.Vinogradova M. Yu. Feigina and N. A. Aldanova Tetrahedron Letters 1964 951 ; N.S.Wulfson V. A. Puchkov B. V. Rozinov Yu. A. Ovchinnikov A. A. Kiryushkin and V. T. Ivanov ibid. 1965,2793. "W. S. Bertand M. C. Probine J. S. Shannon and A. Taylor Tetrahedron 1965,21 677,and references therein. Amino-acids and Peptides on the size of the ring.53-55 Ring-opening by reactions analogous to (18) -+ (19) and (18) + (20) are observed with cyclopeptides and in certain cases other ring-opening reactions have also been observed [e.g. (22) -,(23) and (24) + (25)JS6 The spectra of cyclic peptides and large cyclodepsipeptides (e.g. octa-and deca-depsipeptides) may be impossible to interpret because of the number of positions at which ring opening occurs.CHR'R2 CHR3R4 CHR5R6 I I O-CH-CO-0-CH-CO-0-CH-CO [CO .CHR .O], + CHR'R' CHR3R4 CHR5R6 CHR3R4 I 1 O-CH-CO-b==(!H -CH-CO (20) I 1 CHR' R2 O=C+$i II /CFo/C% R R'CH (21) .+ NH i +[CO CHR NHI [CO .CHR . NH] (22) (23) 0 6 0 NH .CHR' .c& NHKHR [CO .CHR . NH] (24) (25) Staphylomycin S (26)57gives a readily interpretable spectrum [(26) +(27)] because it contains only one ester bond.58 C0,-Type ring opening also pre- dominates for isarolides A (28; R' = Pr' R2 = CH,Ph) B (28; R' = CH2Ph R2 = Pri) and C (28; R' = R2 = CH2Ph) and the structures of these com- " C. H. Hassall and J. 0.Thomas Tetrahedron Letters 1966,4485. " B. J. Millard Tetrahedron Letters 1965 3041. " H. Vanderhaeghe and G. Parmentier J.Amer. Chem. SOC. 1960,82,4414. '* A. A. Kiryushkin V. M. Burikov and B. V. Rosinov Tetrahedron Letters 1967 2675. 460 H.D. Law CO-CH-NH-co 1 Ph CH,Ph -COz -H d I 1 L mle 122 205 290 387 548 673 778 (27) pounds could be postulated on the basis of the fragmentation patterns outlined above even though only a mixture of the compounds was available.59 Fragmentation is not confined to the peptide backbone (vide elimination in the threonine residue in staphylomycin S),but the type of side-chain fragmenta- tion which occurs is well defined.22c Arginine peptides create difficulties owing to their thermal lability but may be converted into ornithine peptides by hydrazinolysis,60 or to pyrimidylornithine derivatives by reaction with 1,3- diketones,60.61 both of which give satisfactory spectra.60 One of the attractions of the mass spectrometric approach to peptide structure is that automatic sequence determination should be possible and the development of com- puterised techniques for interpreting spectral information goes some way towards realising this end.62-64.Peptide Synthesis.-Protecting groups. There is now an almost embarrassing number of protecting groups available for peptide synthesii although synthetic successes to date have utilised relatively few of them. The preferred methods involve benzyloxycarbonyl- t-butoxycarbonyl- and o-nitrophenylsulphenyl- amino-groups; simple alkyl t-butyl benzyl and p-nitrobenzyl esters ;nitro- 59 L. H. Briggs B. J. Fergus and J.S. Shannon Tetrahedron 1966 Suppl. 8 269. 6o M. M. Shemyakin Yu. A. Ovchinnikov E. I. Vinogradova M. U. Feigina A. A. Kiryushkin 61 K. Toi E. Bynum E. Norris and H. A. Itano J. Bid. Chem. 1967 242 1036; T. P. King N. A. Aldanova Yu. B. Alakhov V. M. Lipkin and B. V. Rosinov Experientia 1967,23,428. K. Biemann C. Cone B. R. Webster and G. P. Arsenault J. Amer. Chem. SOC.,1966,88 5598. Biochemistry 1966,5 3454. 63 M. Senn R. Venkataraghavan and F. W. McLafferty J. Amer. Chem. SOC. 1966,88,5593. 64 M. Barber P. Powers,M. J. Wallington and W. A. Woistenholme Nature 1966,212 784. Amino-acids and Peptides 461 tosyl and protonated guanidino-groups; benzyl ethers and thioethers and N'"-benzyl imidazoles. The preparation of t-butoxycarbonyl amino-acids from t-butyl azido-formate has been improved (yields better than 90%) by the use of an auto- titrator to control the pH of the rea~tion.~' Removal of the p-nitrophenol is sometimes difficult if these derivatives are prepared from t-butyl p-nitrophenyl carbonate but this difficulty can be avoided if an equivalent of NN'-dicyclo- hexylcarbodi-imide is added to the mixture of t-butoxycarbonylamino-acid and p-nitrophenol so that the p-nitrophenyl active ester is prepared directly.66 An important development is the use of formic acid to cleave the t-butoxy- carbonyl group.67 The resulting peptide ester formates obtained quantitatively and generally in a non-hygroscopic crystalline form are quite stable except for dipeptide derivatives which tend to form dioxopiperazines when warmed.Formic acid also cleaves N-trityl N-o-nitrophenylsulphenyl and 0-t-butyl ether and ester groups. Further new methods for the removal of the o-nitrophenylsulphenyl group include the use of sulphonic acid imides22e and o-nitrothiophenol.22f These reactions are essentially displacement reactions requiring an excess of the nucleophile. However if the cleavage with benzenesulphonimide takes place at 70" in alcoholic solution only one equivalent of the imide is required and the process is then presumably one of acid-catalysed alcoholysis.2 2e The relative importance of the displacement and alcoholysis reactions will depend upon the nucleophilicity of the anion of the acid employed. With hydrogen chloride for example it is postulated that sulphenyl chloride formation occurs in an initial displacement reaction and the alkyl sulphenate arises from subsequent alcoholysis of this halide.When N-o-nitrophenylsulphenylcysteinederivatives are treated with acids in a variety of solvents and even in alkaline solution N - S migration of the protecting group occurs.68 Protecting groups possessing features of both benzyloxycarbonyl and t-butoxycarbonyl groups have been investigated. It was hoped that acedimethyl- benzyloxycarbonyl and PP-dimethylphenethoxycarbonyl groups would be thermolabile but heating to 130" for 10 min. was required for cleavage.22g The t-pentyloxycarbonyl group has also been extensively ~tudied.~' New protecting groups which formally resemble the benzyloxycarbonyl and t-butoxycarbonyl groups but which are resistant to hydrogen bromide in acetic acid trifluoroacetic acid etc.are the piperidino-oxy~arbonyl~~ and 2,2,2-trichloroethoxycarbony171groups. The piperidino-oxycarbonyl group has been removed by treatment with zinc or sodium dithionate in aqueous acetic acid by electrolytic reduction in a mixture of N-sulphuric acid and tetra- " E. Schnabel Annalen 1967,702 188. " Y. Wolman D. Ladkany and M. Frankel J. Chem. SOC. (C),1967,689. 67 B. Halpern and D. E. Nitecki Tetrahedron Letters 1967 3031. '* I. Phocas C. Yovanidis I. Photaki and L. Zervas J. Chem. SOC.(C) 1967 1506. 69 S. Sakakibara and M. Itoh Bull. Chem. SOC.Japan 1967,40,646. 70 D. Stevenson and G. T. Young Chem. Comm. 1967,900. " T. B. Windholz and D.B. R. Johnston Tetrahedron Letters 1967 2555. 462 H.D. Law hydrofuran (1 1) and by hydrogenolysis over palladised charcoal in aqueous acetic acid. Treatment with zinc in acetic acid at room temperature and heating the derivative under reflux briefly in methanol in the presence of zinc have been used to remove the trichloroethoxycarbonyl group. Hydrogenolysis over palladised charcoal or platinum oxide did not cleave this group. Protection with 2-iodoethoxycarbonyl was investigated a few years ago but did not find favour owing to the toxicity of some intermediate^.^^ Diphenylmethyl esters which are acid-labile do seem potentially useful in general peptide synthesis.22h- 73 Pentamethylbenzyl esters cleaved by treat- ment with trifluoroacetic acid have been advocated for depsipeptide synthesis.74 Phenacyl esters readily prepared from phenacyl bromide are cleaved by hydrogenolysis or by treatment with sodium ben~enethiolate.’~ p-Bromo- phenacyl esters undergo aminolysis too rapidly to be useful in peptide synthesis.75 Problems encountered with side-chain protection of cysteine residues are discussed below in connection with the synthesis of insulin.Presumably symmetrical thioacetals like the acetals developed for hydroxy-group protection in ribonucleotide could be adapted to give the right degree of acid lability for peptide work. An N”-ditrityl arginine derivative obtained surprisingly,~ by treating tri- or tetra-tritylarginine with hydrochloric or hydro- bromic acid might find application.The trityl groups are removed by treatment with hydrogen bromide in acetic acid.22h Formation ofthe peptide bond. Most but not methods described for peptide formation and all approaches to the synthesis of natural peptides have been dependent upon enhancement ofthe reactivity of one of the functional groups of the amino-acid. Usually but not always,77 the carboxy-group of the amino-acid is activated and where possible the activation of peptide terminal carboxy-groups is avoided because of the dangers of racemization. The classical azide route remains important because it does not cause racemization. Otherwise in the most commonly used coupling methods the carboxy-component is activated by conversion into a mixed anhydride or by reaction with NN’-dicyclohexylcarbodi-imide or by making an active ester from it.A recent reinvestigation of the mixed anhydride technique shows that under the most favourable conditions even peptide C-terminal carboxy- groups can be activated with very little ra~emization.~~ Couplings with high molecular weight carbodi-imides have been in~estigated.~~ A disadvantage of the NN‘-dicyclohexylcarbodi-imide technique the isolation of the peptide from the urea by-product should not arise with this new reagent but undesirable ’’ J. Grimshaw J. Chem. SOC.,1965 7136. 73 J. Taylor-Papadimitrion C. Yovanidis A. Paganon and L. Zervas J. Chem. SOC.(C) 1967 1830. 74 F. H. C. Stewart Chem. andZnd. 1967 1960. ” F. H. C. Stewart Austral. J. Chem. 1967,20 787. ” C. B.Reese R. Saffhill and J. E. Sulston J. Amer. Chem. SOC.,1965,89 3366. 77 G. Liisse and W. Giidicke Annalen 1967 100,3314. ’* G. W. Anderson J. E. Zimmerman and F. M. Callahan J. Amer. Chem. Soc. 1967,89 5012. 79 Y. Wolman S. Kivity and M. Frankel Chem. Comrn 1967 629. Amino-acids and Peptides acyl urea formation might prove more troublesome. N N'-Dicyclohexyl-urea can be converted into the carbodi-imide by treatment with phosphorus pentoxide." One drawback to the use of Reagent K (29; R' = Me R2 = SO, R3= H) in peptide synthesis is that the active ester (30)which is formed has a tendency to rearrange to the keto-imide (31;R' = Me R2 = SO; R3= H). N-Ethyl- isoxazolio-derivatives (29; R' = Et R2 = SO, R3= H and 29; R' = Et R2 = H R3 = SO;),'' but not N-aryl isoxazolio-derivatives (29; R' = Ph R2 = R3 = H),'* are better from this point of view.2-Ethyl-7-hydroxy- benzisoxazolium salts (32) react with N-protected amino-acid sodium salts at pH 4-5 to give the active esters (33). These substituted catechol esters are extremely resistant to racemization and can even be reacted with amino-salts to give racemate-free peptides. This reaction seems to offer a means for activating terminal carboxy-groups of peptides without danger of race-mi~ation.*~ The resistance to racemization shown by these esters is undoubtedly due to a variety of factors one of which will be availability of the o-hydroxy-group for hydrogen-bonding with the approaching amine which should facilitate aminolytic attack at the expense of the oxazolone-forming reaction.Similar effects have been observed with esters derived from catechol,22j l-hydroxy- piperidine,22k.84 and 8-hydroxyquinoline.' Esters derived from N-hydroxy- LO -CH .CH .CO .NH .CH * CO .NH * CH .CO .NH .CH .Cd I I I I [CH2I R' R2 CH 1 / '1 Me Me Et R2 NHR~ R3& C@ I Reagents CeL ,N R' .CO CH R"-NHR~ i Rb-NH .CHR"-CO2-(3 1) II 0 ''C. L. Stevens G. H. Singhal and A. B. Ash J. Org. Chem.. 1967,32 2895. " R. B. Woodward R. A. Olofson and H. Mayer Tetrahedron 1966 Suppl. 8,321. R. B. Woodward D. J. Woodman and Y. Kobayashi J. Org. Chem. 1967,32 388. 83 D. S. Kemp and S. W. Chien J. Amer. Chem. SOC.,1967,89,2743. 84 J. H. Jones B. Liberek and G. T. Young J. Chem. SOC.(C) 1967,2371." H.-D. Jakubke A. Voigt and S. Burkhardt Chem. Ber. 1967,100,2367. 464 H. D. Law OH OH 32 \ 0-CO * CHIps NHRb I (3 4) (3 3) succinimide and N-hydroxypiperidine can be prepared by the mixed anhydride method.86 Again under the most favourable conditions acyl peptides seem to couple with little racemization. N-Hydroxyphthalimide p-nitrophenol and 2,4,5-trichlorophenoI do not give active esters by this method whilst penta- chlorophenol and 8-hydroxyquinoline give partially racemized esters. Pivalohydroxamic acid esters (34; R = CMe,) may be prepared by the reaction of acyl amino-acids with pivalonitrile oxide and this approach has also been reported to give no racemization when used to activate a C-terminal carboxy-group [Z.Gly.L.Phe + Gl~.oEt].~~ N-Benzyloxycarbonylproline could not be coupled by this technique because of the prevalence of a side- reaction involving Lossen rearrangement.88 Benzohydroxamic esters (34; R’ = Ph) prepared from the silver salt of the acyl amino-acid or N-protected peptide and benzohydroximic acid chloride are also reported to be free from racemization during their formation and subsequent aminolysis.’ Halogenated phenyl esters especially tri~hloro-,~~. 93 91 penta~hloro-,’~ and pentafl~oro-’~. 94 phenyl esters have received continuing attention. Pentachloro- and pentafluoro-phenol form complexes (35; X = C1 or F) with NN’-dicyclohexylcarbodi-imide which react with acyl peptides to give the corresponding active esters once more with very little ra~emization.’~ How-ever it is thought that the complex dissociates into the phenol and ”’-dicyclohexylcarbodi-imide,and that ester formation takes place in the normal way.Racemization is possibly suppressed because the acidity of the phenol tends to depress the base-catalyzed racemization of the corresponding 86 G. W. Anderson F. M. Callahan and J. E. Zimmerman J. Amer. Chem. SOC. 1967,89 178. S. Rajappp K. Nagarajan and V. S. Iyer Tetrahedron 1967,23,4805. T. R. Govindachari S. Rajappa A. S. Akerkar and V. S. Iyer Tetrahedron 1967 23 4811. E. Taschner B. Rzeszotarska and L. Lubiewska Chem. and Ind. 1967,402. B. Rzeszotarska and G. P. Vlasov Bull. Acad. polon. Sci. SPr. Sci. Chim. 1967 15 143. 91 J. S. Morley J. Chem. SOC. (C) 1967 2410. 92 J.Kovacs M. Q. Ceprini C. A. Dupraz and G. N. Schmidt J. Org. Chem. 1967 32 3696; A. Kapoor and E. J. Davis Experientia 1967 23 253. 93 J. Kovacs L. Kisfaludy and M. Q. Ceprini J. Amer. Chem. SOC. 1967 89 183. 94 V. A. Shibnev T. P. Chuvaeva and K. T. Poroshin Izvest. Akad. Nauk. S.S.S.R.Ser. khim. 1967,954. Amino-acids and Peptides 465 oxazolone which will in any case undergo ring opening more rapidly with these strongly acidic phenols than for example with p-nitrophenol. A new technique for the preparation of p-nitrophenyl esters has been noted above. Synthesis of the elusive Na-benzyloxcarbonyl-N”-nitroarginep-nitrophenyl ester is also reported.” X (35) Amino-acid triethylenetetraminecobalt(111) complexes and the corresponding bisethylenediamine complexes (36) may be used in peptide ~ynthesis.~~ The perchlorate of the bisethylenediamineglycine methyl ester cobalt (111) complex reacts in 1 min.at 20” in ahydrous sulpholan dimethyl sulphoxide or acetone with peptide or amino-acid esters to add a glycine residue to the N-terminal amino-group. Complexing both protects the amino-group of the glycine and apparently activates the ester function. This reaction has not yet been fully studied and further information about racemization yield and generality will be required before its true significance can be estimated. However because of the strict steric requirements of both reactants and products and because of the convenience and consequent efficiency of simultaneous protection and activation it seem’s extremely promising.The technique of ‘solid-phase’ peptide synthesis developed by Mer~-ifield,~~ has made good headway although many laboratories have experienced difficulties when first using the method. In this technique the N-protected C-terminal amino-acid residue reacts as its triethylamine salt with an insoluble chloromethylated polystyrene resin to form a benzyl ester-type link. The N-protecting group is subsequentky removed and N-protected amino-acids are added to the insoluble aminoacyl polymer by a suitable succession of ‘de- blocking’ and ‘coupling’ stages. t-Butoxycarbonyl protection of amino-groups has been used almost exclusively and ”’-dicyclohexylcarbodi-imide has usually been employed in the coupling stage.Cleavage of the peptide from the resin has generally been achieved by the use of hydrogen bromide in trifluoroacetic acid. Quaternization of the chloromethylated resin has been detected as a side 95 W. Godicke and G. Losse 2.Chem. 1967,7,232. 96 D. A. Buckingham G. L. Marzilli and A. M. Sargeson J. Amer. Chem. SOC. 1967 89 2772 4539. 97 R. B. Merrifield J. Amer. Chem. SOC. 1963 85 2149; G. R. Marshall and R. B. Merrifield Biochemistry 1965,4 2394. 466 H.D. Law reaction in the esterification stage."'* 22m Studies in which glycine was bound to the resin as the C-terminal amino-acid show that there is a considerable variation in the reactivity of the various glycine moieties (aminolysis of t- butoxycarbonyl-leucine 5-chloro-8-hydroxyquinoline ester).It has been advo- cated that the first peptide bond-forming stage should be relatively short after which the unreactive sites should be irreversibly blocked to reduce the kinetic heterogeneity of subsequent peptide-forming stages.22' Asparagine and gluta- mine residues are normally introduced as active esters because of the danger of nitrile formation in the presence of NN'-dicyclohexylcarbodi-imide. In recent work on insulin peptides 0-benzyltyrosine S-benzylcysteine and N-benzyloxycarbonyl-lysine have also been incorporated as activated esters.22n 1,2,4-Triazole catalysis is reported to improve yields when active esters are employed in the solid phase technique.22m Removal of the peptide from the resin at the end of the synthesis under strongly acid conditions limits the range of peptides to which solid phase synthesis in its present form can be applied.However a tryptophan residue in a pentapeptide was found to be left intact when the cleavage was carried out with hydrogen chloride in acetic acid.98 Anhydrous hydrogen fluoride9' and hydrazinolysis' O0*'O' have both been used satisfactorily for the cleavage ;hydrolysis has been investigated. lo' Several other techniques for the synthesis of peptides without isolating the intermediates involved (see e.g. Morley' ') have been reported. Peptide synthesis has also been carried out by a 'reverse solid-phase' method in which the activated amino-acid is the insoluble phase.'*" Especially noteworthy is the use of amino-acid N-carboxy-anhydrides in stepwise peptide synthesis.'03 Racemization. The amount of racemization inherent in the use of a particular coupling technique or protecting group has usually been assessed by the synthesis of model di- and tri-peptides. Diastereoisomers resulting have been estimated by recrystallization by t.l.c. or by g.1.c. Three mechanisms oxazolone formation p-elimination and proton withdrawal have been associated with racemization during peptide synthesis and it has been suggested that three tests based on the racemization of the p-nitrophenyl esters of N-benzoyl-L- leucinate N-benzyloxycarbonyl-S-benzyl-L-cysteinateand N-benzyloxycar- bony1 L-phenylglycinate could be used to investigate separately the racemiza- tion occurring by each of these pathway^."^ Differences in the chemical shifts of the methyl protons in diastereoisomeric alanine peptides may be used to assess racemization for example in the for- mation of methyl N-acetyl-L-alanyl-L-phenylalaninate.'O5 The amino-acid 9a A. Loffet Experientia 1967 23,406. 99 J. Lenard and A. B. Robinson J. Amer. Chem. SOC.,1967,89 181. loo M. Ohno and C. B. Anfinsen J. Amer. Chem. SOC. 1967,89 5995. W. Kessler and B. Iselin Helo. Chim. Acta 1966,49 1330. A. Deer Angew. Chem. 1966,78 1064. R. Hirschmann R. G. Strachan H. Schwam L. F. Schoenewaldt H. Joshua J. Barkemeyer D. F. Verber W. J. Paleveda T. A. Jacob T. E. Beesley and R. G. Denkewalter J. Org. Chern. 1967 32 34 15. M. Bodanszky and A. Bodanszky Chem. Comm. 1967,591. 105 B. Halpern L. F.Chew and B. Weinstein J. Amer. Chem. Soc. 1967 89 505; B. Halpern D. E. Nitecki and B. Weinstein Tetrahedron Letters 1967 3075 ; J. 0.Thomas ibid. p. 335. Amino-acids and Peptides 467 analyser is used in another novel racemization test. N-Acetyl-L-isoleucine is coupled to ethyl glycinate and after hydrolysis of the resulting peptide the ratios of isoleucine and alloisoleucine are measured."' A new system for the detection of racemization by g.1.c. lo7 depends upon the coupling of N-protected (t-butoxycarbonyl) amino-acids with an optically pure asymmetric amine [( -)-2-amino-4-methylpentane] and separation of the diastereoisomers pro- duced. Phenylalanine derivatives tend to decompose under the g.1.c. conditions used in this work but the phenylalanine amides can be separated by t.1.c.The general approach has the advantage that coupling methods can be investi- gated for different amino-acids. It is also likely that the N-protecting group could be varied. Racemization during peptide synthesis most commonly occurs via the oxazolone route.2 2j N-Alkoxycarbonyl amino-acids in contrast to N-acyl amino-acids are generally resistant to racemization. Kinetic studies"* 0-I suggest that oxazolone formation occurs via the conjugate base [-N=C-] which might account for these differences. Studies of molecular models show that the amide bond must be in the trans-conformation for oxazolone formation to occur; shifts to more polar solvents tend to stabilise cis-amide bonds. The assistance of the electronic effect favouring oxazolone formation will therefore be offset to some extent in solvents of high polarity by the conformational effect of the solvent on the peptide bond.22p Once formed the oxazolone can undergo the reverse reaction to give the unchanged activated carboxy-derivative or can racemize and then be converted into the racemic starting material.The oxazolone before or after racemization can also react with the amine com- ponent to give the peptide directly. In these circumstances it is not surprising that the oxazolone may be difficult to detect even when it can be clearly demonstrated by ancillary evidence that racemization is proceeding via this mechanism."' Generally it may be taken that racemization of the oxazolone will be at least as fast and generally faster than the various ring opening reactions which the oxazolone might undergo although the rate of ring opening does vary significantly with the strength of the nucleophile involved.110 The problem of coupling acyl peptides to amino-components without racemization is therefore the problem of activating the carboxy-group towards aminolytic attack without activating it for oxazolone formation.This has been one of the outstanding problems of modern synthetic peptide chemistry and the continuing importance of the azide coupling technique is directly attributable to its success in achieving this end. The development of new techniques which might achieve racemization-free coupling of this type has lo6 M. Bodansky and L. E. Conklin Chem.Comm. 1967 773. lo' B. Halpern L. F. Chew and J. W. Westley Anafyt.Chem. 1967,39 399. D. S. Kemp and S. W. Chien J. Amer. Chem. SOC.,1967,8!4,2745. lo9 I. Antonovics and G. T. Young J. Chem. SOC.(C),1967,595. W. H. McGahren and M. Goodman Tetrahedron 1967 23 2017; M. Goodman and W. H. McGahren ibid. p. 2031. 468 H. D. Law been noted in this review. These techniques if generally applicable are likely to prove very important. Natural Peptides.-All of the natural peptides so far synthesized were first prepared by what may be regarded as the classical route which involves the isolation and where possible the characterization of all intermediates. Approaches of this kind though unequivocal would be prohibitively laborious and inefficient in the synthesis of very large compounds and unless new methods of isolation are introduced which are less dependent on fortuitous physical properties and on the chemists' manipulative skill it seems possible that insulin 111-113 will remain the largest peptide to be synthesized by this method.The alteI-.ative is to employ a route in which the isolation of intermediates is reduced to a minimum. Thanks to chromatographic and instrumental tech- niques this does not mean that characterization must be neglected. Ideally the technique should be susceptible to automation. Merrifield's solid-phase approach is the most thoroughly investigated technique of this kind and by its use variously protected A-and B-chains of insulin have been prepared in high yields in a matter of 22q* l4 Th e yield of insulin when the synthetic chains prepared in this way are combined is not as large as the yield obtained by recombination of the chains isolated from the natural hormone.However these differences seem to be due to the limitations of the protecting groups used and not to a conceptual failing of the solid-phase approach. Experiments in which o-amino-acids were incorporated into polyamides suggest that the synthesis of peptides at least twice the length of the insulin chains should offer no new fundamental difficulties.22' Several peptides smaller than insulin have been prepared by the solid-phase approach and they have generally been obtained more conveniently and in higher yield than by the classical route.'15 At least two major difficulties still beset the would-be synthesizer of insulin.The first arises from the fact that the molecule contains three disulphide bonds so that an unequivocal synthesis would require three types of sulphur pro- tecting groups with different labilities to enable the individual disulphide bonds to be made unambiguously one at a time. The second is the protecting group difficulty mentioned above. In all insulin syntheses reported to date cysteine residues have been incorporated as the benzyl thioether deriva- tive.' 11-113* Although the use of only one S-protecting group could not be expected to given an unambiguous combination of the two chains at the J. Meienhofer. E. Schnabel H. Bremer 0. Brinkhoff. R. Zabel. W. Sroka H. Klostermeyer.D. Brandenburg T. Okuda and H. Zahn Z. Naturforsch. 1963 18b. 1130. '12 P. G. Katsoyannis K. Fukuda A. Tometsko K. Suzuki and M. Tilak J. Arner. Chem. SOC. 1964,86,930. K. T. Kung Y. C. Du W. T. Huang C. C. Chen L. T. Ke S. C. Hu R. Q. Jiang S. Q. Chu C. I. Niu J. Z. Hsu W. C. Chang L. L. Chen H. S. Li Y. Wang T. P. Loh A. H. Chi C. H. Li P. T. Shi Y. H. Yieh K. L. Tang and C. Y. Hsing Sci. Sinica 1965 14 1710. 'I4 A. Marglin and R. B. Merrifield J. Amer. Chem. SOC.,1966,88 5051. G. R. Marshall and R. B. Merrifield Biochemistry 1965 4 2394; J. M. Stewart and D. W. Woolley Fed. Proc. 1965 24 657; M. C. Khosla R. R. Smeby and F. M. Bumpus Biochemistry 1967 6 754; W. K. Park R. R. Smeby and F. M. Bumpus ibid. p. 3458. '16 H. Klostermeyer and R.E. Humbel Angew. Chem. 1966,78 871 ; P. G. Katsoyannis Science 1966,154 1509. Amino-acids and Peptides conclusion of the synthesis studies with the natural chains show that a con- siderable amount of insulin activity can be obtained by recombination of the chains under appropriate conditions.' ' Sodium-liquid-ammonia reduction is necessary for the cleavage of the S-benzyl group and to remove other protecting groups N"-tosyl and N'"-benzyl used in these syntheses. This treatment has created problems in the synthesis of small peptides but these have not been insuperablq' '*whereas in the insulin work the reductive step leads to exten- sive side reactions including desulphuration' l9 and splitting of the peptide chain.'20 These side reactions account for the consistently low yields of insulin activity when recombination data for natural and synthetic chains are com- pared.An attempted synthesis of the B-chain of human insulin failed because of the destruction during the sodium-liquid ammonia treatment of the C-terminal threonine residue. ' Syntheses which employ other protecting groups are therefore required. Another approach using cystinyl peptides has recently been reported.'22 The N-benzyloxycarbonyl hexadecapeptide symmetrical disulphide possessing the residue sequence of the N-terminal portion of the B-chain was combined by the azide technique with the tetradecapeptide symmetrical disulphide equivalent to the C-terminal sequence. t-Butyl protecting groups were used on side chains and for the C-terminal carboxy-group ; arginine guanidino- groups were protonated.The main product of the coupling was a polymeric material possessing the full 1-30 amino-acid sequence. Protecting groups could be removed and the product could be cleaved by sulphitolysis to give the B-chain 17,19 S-sulphonate in 2-4 % yield based on the crude polymers. When combined with natural A-chain this material yielded 1-2.6 I.U./mg. of insulin activity whereas the B-chain after sodium-liquid ammonia reduction gave 0.5-0.75 I.U./mg. In the same experiments recombination of natural A- and B-chain S-sulphonates gave 2-3 I.U./mg. The attempt to prepare insulin by an unambiguous route employing three sulphur-protecting groups and forming each disulphide bridge in turn has the inherent difficulty that unsymmetrical cystine peptides tend to be unstable.Clearly formed disulphide bond(s) need to remain intact whilst the other protecting groups are removed and the new disulphide bond(s) formed. Several studies of the synthesis and stability of unsymmetrical open-chain cystine peptides have been reported in recent years.22h' 123,124 S-Trityl S-11' P. G. Katsoyannis A. Tometsko C. Zalut S. Johnson and A. C. Trakatellis Biochemistry 1967 6 2635; P. G. Katsoyannis A. C. Trakatellis S. Johnson C. Zalut and G. Schwartz ibid. p. 2642 and references therein. H. Nesvadba and H. Roth Monatsh. 1967,98 1432. l9 P. G. Katsoyannis Amer. J. Med. 1966,40 652. W. F. Benisek and R. D. Cole Biochem. Biophys. Res. Comm. 1965 20 655.lZ1 H. Zahn T. Okuda and Y. Shimonishi Angew. Chem. 1967,79,424. 12' H. Zahn and G. Schmidt Tetrahedron Letters 1967 5095; H. Zahn and W. Sroka Annalen 1967,706,230. 12' R. G. Hiskey and E. L. Smithwick jun. J. Amer. Chem. SOC. 1967,89,437. lZ4 R. G.Hiskey T. Mizoguichi and E. L. Smithwick jun.,J. Org. Chem. 1967,32,97 ;R. G. Hiskey J. T. Staples and R. L. Smith ihid. p. 2772 R. G. Hiskey and M. A. Harpold Tetrahedron 1967,23 3923 and references therein. 470 H.D.Law diphenylmethyl and S-acyl cysteine derivatives appear to offer the possibility of selective removal. The question of the stability of the disulphide is not so clear-cut but stability of unsymmetrical disulphides under acidic conditions has been demonstrated. For example N-diphenylmethoxycarbonyl protecting groups and t-butyl esters may be cleaved in unsymmetrical cystine peptides with boron trifluoride in acetic acid without disturbing the di~ulphide.'~~ S-Trityl groups are stable under these conditions.Ring closure of unsym-metrical cystine peptides by peptide bond formation without disruption of the disulphide has also been described but is not always satisfactory.'23-125 S03H I Elu. Gln. Asp. Tyr. Thr. Gly. Try. Met . Asp. Phe. NH 12 3 4 5 6 7 8 9 10 Isolation and structural and synthetic investigations of a new peptide caerulin (37) have been announced.'26 Caerulin which occurs in the skin of Hyla caerulea acts on vascular and extravascular smooth muscle and on external secretions. That it stimulates gastric secretion is not surprising since the C-terminal pentapeptide amides of caerulein and gastrin are the same.It has been amply demonstrated that the C-terminal tetrapeptide sequence in gastrin is all that is required for biological activity to be manifest.'27 However whereas the free phenol of gastrin and the sulphated material both occur naturally and are equally active in stimulating gastric secretion the caerulin isolated is fully sulphated and the free phenol of this peptide is not as active as the sulphate. The evolutionary significance of the similarities between these two peptides is obscure. Structures of gastrins from pig (38; A = Met B = C = Glu) human (38; A = Leu B = C = Glu) sheep (38; A = Val B = Glu C = Ala) cattle (38; A = Val B = Glu C = Ala) and dog (38; A = Met B = Ala C = Glu) are known;22s syntheses of human gastrin have recently been reported.22'* 128 919 S03H I Gh.Gly .Pro. Try. A. Glu . B. Glu . C . Ala . Tyr . Gly .Try. Met . Asp. Phe .NH 1 2 3 4 5 6-7 8 9 10 11 12 13 14 15 16 17 I. Photaki J. Amer. Chem. SOC.,1966,88 2292. D. Anastasi V. Erspamer and R. Endean Experientia 1967,23 699; L. Bernardi G. Bosisio R. de Castiglione and 0.Goffredo ibid. p. 700; V.Erspamer G. Bertaccini G. de Caro R. Endean and M. Impicciatore ibid. p. 702. J. M. Davey A. H. Laird and J. S. Morley J. Chem. SOC.(C) 1966,555 and references therein. J. Beacham P. H. Bentley G. W. Kenner J. K. MacLeod J. J. Mendive and R. C. Sheppard J. Chem. SOC.(C) 1967,2520. Amino-acids and Peptides 47 1 Standard nomenclature for analogues of natural peptides has been formu- lated.' 29 Recent studies'30* '31 of analogues of the ribonuclease S-peptide are particularly interesting because of the similarities between the activation of S-protein by S-peptide and peptide-hormone action.'Binding-sites' res-ponsible for the attachment of the S-peptide to the S-protein involve the 'Glu 14Asp and 3Met residues. These residues are not essential for biological activity although they do exert an effect. On the other hand the histidine residue in the 12-position is essential for activity. P-(Pyrazol-3-yl)-alanine is sterically similar to histidine but its acid-base properties are quite different.I3O The analogue of the 1-14 residue S-peptide containing this amino-acid residue at position 12 instead of histidine does not activate the S-protein even when present in a molar ratio of 10oO :1 but it is a potent competitive inhibitor of S-peptide a~tivation.'~' 0.r.d.studies suggest that the S-peptide has a helical configura- tion.' 32 Staphylococcal nuclease which can be cleaved into two peptides which are enzymically active when mixed although not covalently bonded provides a new system for similar studies of peptide interaction^.'^^ Cyclic Peptide.-Notable degradative studies of the amanitine cyclo- octapeptides (39; a R' = NH, R2 = OH; b R' = OH R2= OH; c R' = NH, RZ = H) derived from Arnanita phalloides have been described in full.' CH,R H a +OH Me-CH I NH-C)1-CO-NH-CH-CO-NH-CH2-C0 I I I co I Hq CILo~~ T:- CH ,Me Et HO co XH2 H I CO-CH-NHO-~I-I-NltC~CH2-NH I CH .CO.R' (3 9) Dimerisation during the cyclisation of linear peptides is well known.'34 It was originally observed during the synthesis of gramicidin S cyclo-(~-Val.~-Orn.~-Leu.~-Phe.~-Pro), via the cyclisation in pyridine solution of the IUPAC-IUB Commission on Biochemical Nomenclature Biochemistry 1967,6 362. F. M. Finn and K. Hofmann J. Amer. Chem. Soc. 1967,89,5298; K. Hofmann and H. Bohn ibid. 1966,88 5914 and references therein. 13' F. Marchiori R. Rocchi G. Vidali A. Tamburro and E. Scoffone J. Chem. SOC.(C),1967 81 ; R. Rocchi F. Marchiori A. Scatturin and E. Scoffone ibid.,p. 86; F. Marchiori R. Rocchi L. Moroder G. Vidali and E. Scoffone ibid.p. 89; E. Scoffone R. Rocchi F. Marchiori A. Marzotto A. Scatturin A. Tamburro and G. Vidali ibid. p. 606; E. Scoffone R. Rocchi F. Marchiori L. Maroder A. Marzotto and A. M. Tamburro J. Amer. Chem. SOC. 1967,89 5450. 132 E. Scoffone at the Chemical Society Anniversary Meetings Exeter April 1967. 133 H. Taniuchi C. B. Anfinsen and A. Sodja Proc. Nut. Acad. Sci. U.S.A. 1967,58 1235. 134 R. Schwyzer and P. Sieber Helv. Chim. Acta 1958,41,2186. 2190 2199. 472 H.D. Law p-nitrophenyl ester of Val.(Tos)Orn.Leu.D-Phe.Pro and was attributed to association between two pentapeptide moieties.' 34 The corresponding cyclo- pentapeptide cyclosemigramicidin S has now been obtained together with the cyclodecapeptide from a similar cyclization of the N-benzyloxycarbonyl- ornithine deri~ative.'~' Cyclosemigramicidin S is not active against a spectrum of micro-organisms tested.Retroenanti~-gly~~'~-gramicidin S which is like gly5."-gramicidin S and differs from it only in the direction of the amide bonds has antibacterial activity similar to that of the parent compound.'36 Sometimes dimerisations occur when peptides which cannot associate by hydrogen bonding are cyclised e.g. di-L-prolylglycine.22" In these cases strain or hindrance in the transition states which lead to the cyclotripeptides must account for the dimerisations. Penta- and hexa-peptides are reported to cyclise readily on treatment with o-phenylene pyrophosphite and very little cyclodimerization occurs.' 37 Other rings with less than eighteen atoms and containing N"-amino-acids can be prepared by aminoacyl insertion reactions e.g.[(40)-,(41)].22"Hydroxyacyl insertion reactions have been observed in cyclodepsipeptides and it has been postulated that biosynthesis of cyclo- depsipeptides might occur in this way. ' '* Cyclodimerisation with depsipep- tides can also be demonstrated however [e.g. (42) -+ (43; R = CHMe or OBu' or OH) and must be considered a biosynthetic possibility. The serine- containing cyclodepsipeptides (43; R = OBu' or OH) are in this case obtained in two forms which are possibly configurational isomers although the two forms (43; R = OH) can be sublimed unchanged.'39 OH (40) CH2-NH+O-CH2 / \ c\Hz /CH2 CH2-CO-NH-CH2 lJ5 M. Waki and N. Izumiya J. Amer.Chem. SOC.,1967,89 1278. 136 M. M. Shemyakin Yu. A. Ovchinnikov V. T. Ivanov and I. D. Ryabova Experientia 1967 23 326. 13' A. W. Miller and P. W. G. Smith J. Chem. SOC.(C),1967 2140. V. K. Antonov A. M. Shkrob V. I. Shchelokov and Z. E. Agadzhanyan Tetrahedron 1965 21,3537; D. W. Russell Quart. Rev. 1966,20 556. lJ9 C. H. Hassall T. G. Martin J. A. Schotield and J. 0.Thomas J. Chem. SOC.(C),1967 997. Amino-acids and Peptides 473 CH,.R I P NH CH .CO-O*CH,- CH,.CO,H yH2 \ ,NH-~H-co-o, (42) Yo c\H2 CH FH2 \ CH co ‘0 -CO-CH-NH/ (43) kH2 ‘R Peptide Conformation.-The conformations occurring in fibrous proteins and poly-ct-amino-acids and the fundamental importance of helical random coil and pleated sheet structures have been understood for some years.14’ Since poly-L-proline exists in helical forms,141 helix formation cannot be dependent on hydrogen bonding and this is further borne out by studies of poly-(N-methyl-L-alanine) which exists as a helix despite the absence of hydrogen bonds and pyrrolidine rings.’42 In conformational work interactions between side chains and between side chains and solvent molecules have to be taken into account as well as hydrogen-bonding possibilities.Recently the total configurations of several globular proteins have been determined by X-ray crystallography. 143 There is some indication of the relationship between these configurations determined for the solid state protein and the configura- tions of the proteins in solution but no methods comparable to X-ray crystallography are available for the direct determination of the configurations of proteins in solution.The best techniques available give parameters which may be interpreted in terms of postulated structures. In particular 0.r.d. and c.d. have been much used to gain an insight into the secondary structure (i.e. helix uersus random coil) of the peptide backbone. These methods give no indication of the overall three-dimensional structures of proteins because tertiary structure i.e. the folding of the helix or random coil is not taken into account. Solution experiments only give an indication of the general shape of the mole- cule; thin film dialysis’44 and tritium-hydrogen exchange145 are examples of recently developed techniques of this type.140 W. F. Harrington R. Josephs and D. M. Segal Anc. Rev. Biochem. 1966 35 599; S. N. Timasheff and M. J. Gorbunoff ibid. 1967,36 13. 14’ W. Traube and U. Shmueli Nature 1963 198 1165; P. M. Cowan and S. McGavin ibid. 1955 176;’501 ;J. Engel Biopolymers 1966,4 945. 14’ M. Goodman and M. Fried J. Amer. Chem. SOC.,1967,89,1264;J. E. Mark and M. Goodman ibid. p. 1267. 143 J. C. Kendrew H. C. Watson B. E. Strandberg R. E. Dickerson D. C. Phillips and V. C. Shore Nature 1961,190 663; C.C. F. Blake D. F. Koenig G. A. Mair A. C. T. North D. C. Phillips and V. R. Sarma ibid. 1965 206 757; D. Karthra J. Bello and D. Harker ibid. 1967 213 862; H. P. Avey M. 0.Boles C. H. Carlisle S. A. Evans S. J. Morris R. A. Palmer B. A. Woolhouse and S.Shall Nature 1967 213 557. 144 L. C. Craig E. J. Harfenist and A. L. Paladini Biochemistry 1964 3 764; M. A. Ruttenberg T. P. King and L. C. Craig ibid. 1966 5 2857. 14’ M. Saunders H. A. Jung and W. L. Hamilton J. Amer. Chem. SOC. 1967,89,472. 474 H. D. Law Important information on the effectiveness of the physical methods employed and on the basic structures involved has been derived from experiments with mixed solvents. In some solvents (e.g.deuteriochloroform) poly-z-amino-acids tend to exist in helical configurations whilst in other solvents (e.g. trifluoro-acetic acid) the random coil configuration is preferred. At intermediate solvent compositions varying proportions of helix and random coil exist and trans- formation from one form into the other can be demonstrated by changing the solvent proportions accordingly.Solvent interactions might be expected to contribute to the conformation-regulating properties of intermediate mix- ture~.'~~ Protonation of the amide'47 and hydrogen bonding of the trifluoro- acetic acid to the peptide -NH-I4* have been suggested as first steps in helix + random coil transitions but other studies have provided no support for these mechanism^.'^^ In suitable cases random coil and helical structures can be distinguished by n.m.r. studies,'" although proteins are too complex for detailed analy~is.'~ The random coil C(a)H resonance occurs downfield from the helix C(a)H peak and the random coil NH peak up-field from the corresponding helix peak. This applies to both left hand (poly-L-leucine) and right hand [poly-p- methyl-L-aspartate)] helices.149 The helix content indicated by high resolution n.m.r studies seems generally to correlate well with the helix content calculated from 0.r.d. data.149.150 but all authors do not agree on this point.'48 N.m.r. is potentially useful for conformational studies on small peptides because it also gives proton signals generated by the side chains of the residues. Optical techniques have also been used for the study of conformation in small peptides but interpretation of the results can be difficult.152 0.r.d. measurements indicate that helical structures are not important in oligopeptides of y-benzyl-L-glutamate below the pentamer-nonamer range (in dimethylformamide or rn-cresol) nor in copolymers of L-alanine and y-benzyl-L-glutamate below the cononamer (in trifluoroethanol) and this is undoubtedly because insuficient intramolecular hydrogen bonding exists in the smaller peptides to stabilise the ~t-he1ix.I~~ However it is probable that small peptides do not have random configurations.Ethyl t-butoxycarbonyl-L- valyl-L-valyl-L-alanylglycinate for example shows quite definite evidence of C.-C. W. Chao A. Veis and F. Jacobs J. Amer. Cnem. SOC.,1967,89,2219. 14' S. Hanlon Biochemistry 1966,5 2049. 148 W. E. Stewart L. Mandelkern and R. E. Glick Biochemistry 1967,6 143 150. J. A. Ferretti Chem. Comm. 1967 1030. M. Goodman and Y. Masuda Biopolymers 1964 2 107; D. I. Marlborough K. G. Orrell and H. N. Rydon Chem.Comm. 1965,518;J. L. Markey D. H. Headows and 0.Jardetsky J. MoZ. BioZ. 1967 27 25; E. M. Bradbury C. Crane-Robinson and H. W. E. Rattle Nature 1967,216 862. C. C. McDonald and W. D. Phillips J. Amer. Chem. SOC. 1967,89 6332; J. H. Bradbury and H. A. Scheraga ibid. 1966,88,4240; D. P. Hollis G. McDonald and R. L. Biltoneu Proc. Nat. Acd. Sci. U.S.A. 1967 58 758. A. F. Beecham Tetrahedron Letters 1967,211 ;P. M. Scopes D. R. Sparrow J. Beacham and V. T. Ivanov J. Chem. SOC. (C),1967,221 and references therein. M. Goodman E. E. Schmidt and D. A. Yphantis J. Amer. Chem. Soc. 1962 84 1288; M. Goodman M. Langsam and I. G. Rosen Biopolymers 1966,4,305. 14' Amino-acids and Peptides 475 secondary structure in appropriate solvents.' 54 Dielectric constant measure- ments of aqueous solutions of small peptides show that only a very small proportion of the possible conformers of these peptides can make major contributions to the conformer population.'55 In the case of L-alanine tri- and tetra-peptides the conformations making major contributiogs possess angles' 56 suggesting an elongated conformation intermediate between the extended chain and the P-pleated sheet structure.The similai ity between serine and alanine peptides indicates that these conformational preferences do not arise as a result of the hydrophilic-hydrophobic properties of the side The importance of non-bonded interactions of side chains is well illustrated by simple dipeptides in which considerations of this type alone limit the major contributing conformers to 52 % for glycylglycine 16 % for glycyl-L-alanine and 4.5 % for glycyl-L-valine of the total possible conformer population.Branching beyond the y-carbon atom imposes few further restrictions on the possible conformations. ' For larger peptides viz. oxytocin and vasopressin cal- culations based on a simplified expression for the energy of a polypeptide in aqueous solution suggest that a number of minimum energy conformations are available to the peptide.'57b Non-equivalence of the methylene group protons in the peptide backbone has been reported for glycine-containing dipeptides. It is attributed to rigidity of conformation in the dipeptide.15* The lower-field methyl doublet in LL-or DD-as opposed to DL-or LD-alanine dipeptides presumably arises because of the shielding of the methyl group by the adjacent side chain"* and as expected this is particularly marked when the shielding side-chain is aromatic.Similar differences have been reported in other dipeptides and the conformational implications discussed.' 59 Whereas the LL- and the DD-seem to exist in open form the m-dipeptides are more compact. These conclusions are also borne out by the study of dielectric constants the charge separation in all -L-or all -D-forms is consistently greater than in the corresponding diastereoisomers,' and in cyclisation experiments in which peptides and depsipeptides composed of D-and L-residues give higher yields of cyclic product than the all-L- or ali-~-isomers.'~~~ The protons on the CL-and P-carbons of aromatic residues lS4 J.E. Shields and S. T. McDowell J. Amer. Chem. SOC., 1967,89 2499. P. M. Hardy G. W. Kenner and R. C. Sheppard Tetraheron 1963 19 95; J. Beacham V. T. Ivanov G. W. Kenner and R. C. Sheppard Chem. Comm. 1965,386. J. T. Edsall P. J. Flory J. C. Kendrew A. M. Liquori G. Ntmethy G. N. Ramachandran and H. A. Scheraga Biopolymers 1966,4 130; J. Biol. Chem. 1966,241 1004; J. Mol. Biol. 1966 15 339. 15' (a)S. J. Leach G. Ntmethy and H. A. Scheraga Biopolymers 1966,4,369;see also D. A. Brant and P. J. Flory J. Amer. Chem. Soc. 1965 87 2791; P. de Santis E. Giglio A. M. Liquori and A. Ripamonti Nature 1965 206 456; (b) K. D. Gibson and H. A. Scheraga Proc. Nat. Acad. Sci. U.S.A.,1967,58 420 151 7. "13 J.W. Westley and B. Weinstein Chem. Comm. 1967 1232. 159 T. Wieland and H. Bende Chem. Ber. 1965,98 504; F. A. Bovey and G. V. D. Tiers J. Amer. Chem. SOC., 1959,81,2870. 160 G. W. Kenner P. J. Thomson and J. M. Turner J. Chem. SOC. 1958,4148;Yu. A. Ovchinnikov, V.T. Ivanov A. A. Kiryushkin and M. M. Shemyakin Bull. Acad. Sci. U.S.S.R.(Div. Chem. Sci.) 1963,153,1342. 476 H. D.Law in dioxopiperazines are shifted to high field suggesting that the arylmethyl side-chain faces the dioxopiperazine ring. l6 ' The difficulty of determining the conformations of peptides in solution is well illustrated by the case of gramicidin S. Theoreticai considerations suggest that the peptide has an antiparallel intramolecular P-sheet structure with two hydrogen bonds'62 or that it consists of two single turns of a right handed helix.'63 An antiparallel P-sheet structure with four hydrogen bonds was originally favoured by X-ray evidence,164 and because it could easily be derived from the intermolecularly hydrogen-bonded anti parallel arrangement of the linear pentapeptide molecules postulated to account for the cyclodimeri- zation reaction.'65 A third proposed structure packs like an a-helix with four intramolecular hydrogen bonds but consists of three regions; in the first (Val-Om) and third (Leu-Phe) regions the N-H bonds point in the same direction perpendicular to the plane of the ring; in the second (Orn-Leu) the N-H bonds point in opposite ways.164 Other structures have been proposed including one in which the importance of hydrophobic-hydrophilic forces is considered paramount.'66 This model can be ruled out because it involves all-cis-amide bonds whereas it is probable that the amide bonds in gramicidin S are all in the energetically more favourable trans-c~nformation.'~~ The formulation of this structure does emphasize the present difficulty of assessing the importance of hydrophilic-hydrophobic forces.Measurements involving U.V.and i.r. spectroscopy peptide hydrogen-deuterium exchange o.r.d. c.d. dialysis and surface st~dies'~' indicate that none of the proposed models is entirely sati~factory.'~~ These studies also indicate that it might be misleading to assume that only one type of structure can give rise to an a-helix type 0.r.d. pattern.'67 It is clear that new methods are required particularly for studies of peptide conformations in aqueous solution.An interesting development involves the detection and measurement of the effect exerted by one side chain on another non-bonded side chain in its vicinity. Thus 2-bromoacetamido-4-nitrophenol reacts with chymotrypsin to form a sulphonium salt at the essential methionine residue in position 192. The pH dependence of the absorption spectrum of the modified protein suggests that a positively charged imidazole ring is spacially close to the hydroxy-function of the 'reporter group' although there is not a histidine residue near to the methionine residue in the primary structure of the molecule. The reaction of chymotrypsin with ~-bromo-4-nitroacetophenone has also been studied.Monoalkylation at methionine 192 occurs and the alkylated compound exhibits a new absorption peak at A,, 350 mp. This peak 16' K. D. Kopple and D. H. Marr J. Arner. Chem. SOC.,1967,89,6193. G. Vanderkooi S. J. Leach G. Nernethy and H. A. Scheraga Biochemistry 1966,5,2991. 163 A. M. Liquori P. de Santis A. L. Kovacs. and L. Mazzarella Noture. 1966.211. 1039 164 D. C. Hodgkin and B. M. Oughton Biochern. J. 1957,65 752. R. Schwyzer Rec. Chem. Progr. 1959,20 147. D. T. Warner Nature 1961,190 120. 16' D. Balasubrarnanian J. Arner. Chem. SOC.,1967,89 5445. M. B. Hille and D. E. Koshland jun. J. Arner. Chem. SOC., 1967,89 5945. Amino-acids and Peptides is due to the formation of a charge-transfer complex with the indole moiety of a tryptophan residue.It follows that the methionine sulphur is at a distance <SA from the centre of an indole ring. The charge-transfer peak can be reversibly destroyed by denaturation in 8M-urea at pH 3.0.16' A systematic study is in progress to find charge-transfer donors and acceptors suitable for peptide work and preliminary investigations with simple inter- and intra- molecular charge-transfer complexes of amino-acid derivatives have been described.22" One possible criticism of this general approach is that the conformation of the modified protein might be different to that of the un- substituted compound. 169 D. S. Sigman and E. R. Blout J. Amer. Chem. SOC.,1967,89,1747.